ANNUAL ADDRESS OF THE PRESIDENT. XLI 
improve architecture. Similarly, the multiplication of minor inves- 
tigations improves our knowledge of details, but rarely affects the 
great philosophic theories upon which science is founded. The im- 
portance of human actions is measured by the degree in which they 
affect human thought, and the only way of permanently affecting 
scientific thought is by modifying or extending scientific theories. 
The men who do that are neither numerous, nor do they require 
weekly paragraphs to record their deeds; but their names are hon- 
ored by posterity. Even in this golden age the advance of science 
is not steady, but is made by spasmodic leaps and bounds. Mere 
scientific brick making, commonly called progress, is always the 
order of the day until some genius startles the world by a discovery 
affecting accepted theories. Then every effort is directed in the 
new line of thought until it is measurably worked out, and after 
that brick making again resumes its place. While the progress in 
two centuries has been immense, the progress in a week or a month 
is usually almost nil. Optimism has its uses in many departments 
of human affairs, but science should be cool and dispassionate, 
having regard only for the truth. To make a trustworthy estimate 
of the actual state of the whole vast realm of science would be a 
task beyond the powers of any one man; but perhaps it will not 
be amiss to spend the time at our disposal this evening in briefly 
reviewing the recent progress and present condition of the funda- 
mental processes upon which the exact sciences rest—I allude to 
the methods of weighing and measuring. 
Physical science deals with many quantities, but they are all so 
related to each other that almost every one of them can be ex- 
pressed in terms of three fundamental units. As several systems 
of such units are possible, it is important to select the most con- 
venient, and the considerations which guide us in that respect are 
the following: 
1. The quantities selected should admit of very accurate com- 
parison with other quantities of the same kind. 
2. Such comparisons should be possible at all times, and in all 
places. 
3. The processes necessary for making such comparisons should 
be easy and direct. 
